Should Investors Trust Morningstar Ratings? An empirical study on Morningstar ratings ability to predict future performance using evidence from mutual funds in the U.S.

Abstract

This thesis studies the ability Morningstar ratings have to predict future performance, particularly if funds that are top rated tend to outperform the worst rated funds. Our study is focused on the U.S large-cap mutual fund market, and is based on two panel data regression models that use different performance metrics. Specifically, we use three factor models to measure performance; CAPM, Fama-French 3-factor and Carhart 4-factor, to calculate alphas for all funds in the sample over different time horizons. S&P500 is used as benchmark index and we use factor returns developed by Kenneth French. We then use two different panel data regressions, where we regress the alphas against dummies for each rating. The first regression utilizes simple monthly alphas and lagged dummies, whereas the second regression is based on alphas over different time horizons. We find that low Morningstar ratings generally indicate poor future performance. This holds consistently for both analyses we conduct. For investment periods of 6, 12 and 36 months, low rated funds generally underperformed the higher rated funds. We also find that 5-star rated funds fail to consistently outperform the 4- and 3-star rated funds. Based on our findings, it can be concluded that the ratings are more effective in identifying potentially poor investments rather than potentially good, or at least, great ones. Therefore, while the ratings do possess some informational value, they should not be the sole determinant when making investment decisions in mutual funds. The findings are consistent with previous research conducted by Blake and Morey (2000) and Morey and Gottesman (2006), although with stronger evidence of the underperformance of low rated funds.